arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2203.06382cs.CV
收藏

用于基于组的指代表达理解的差异化相关性嵌入

Differentiated Relevances Embedding for Group-based Referring Expression Comprehension

The Hong Kong Polytechnic University(香港理工大学);University of Glasgow(格拉斯哥大学);Xiamen University(厦门大学);Tsinghua University(清华大学);Huawei Noahs Ark Lab(华为诺亚方舟实验室);Fuzhou Institute for Data Technology(福州数据技术研究院)

Fuhai Chen, Xuri Ge, Xiaoshuai Sun, Yue Gao, Jianzhuang Liu, Fufeng Chen, Wenjie Li

AI总结:

针对基于组的REC中多属性、多值导致的负样本爆炸与相关性差异问题,提出多组自步相关性学习及组间约束,提升跨模态匹配性能。

摘要:

指代表达理解的关键在于捕捉跨模态的视觉-语言相关性。现有工作通常在每张图像中对跨模态相关性进行建模,其中锚定对象/表达及其正样本表达/对象与负样本表达/对象具有相同的属性,但属性值不同。这些对象/表达被专门用于通过一对不同的值来学习属性的隐式表示;然而,由于每个锚定对象/表达通常具有多个属性,而每个属性通常具有多个潜在值,这会妨碍属性表示、表达/对象表示及其跨模态相关性的准确性。为此,我们研究了一个名为基于组的REC的新型REC问题,其中每个对象/表达同时用于在语义相似的图像之间构建多个三元组。为了解决负样本爆炸以及锚定-负样本相关性分数差异化的问题,我们提出了多组自步相关性学习机制,根据组内对象-表达对的跨模态相关性,自适应地为其分配不同的优先级。由于不同组之间的平均跨模态相关性差异很大,我们进一步设计了组间相关性约束,以平衡组优先级的偏差。在三个标准REC基准上的实验证明了我们方法的有效性和优越性。

英文摘要:

The key of referring expression comprehension lies in capturing the cross-modal visual-linguistic relevance. Existing works typically model the cross-modal relevance in each image, where the anchor object/expression and their positive expression/object have the same attribute as the negative expression/object, but with different attribute values. These objects/expressions are exclusively utilized to learn the implicit representation of the attribute by a pair of different values, which however impedes the accuracies of the attribute representations, expression/object representations, and their cross-modal relevances since each anchor object/expression usually has multiple attributes while each attribute usually has multiple potential values. To this end, we investigate a novel REC problem named Group-based REC, where each object/expression is simultaneously employed to construct the multiple triplets among the semantically similar images. To tackle the explosion of the negatives and the differentiation of the anchor-negative relevance scores, we propose the multi-group self-paced relevance learning schema to adaptively assign within-group object-expression pairs with different priorities based on their cross-modal relevances. Since the average cross-modal relevance varies a lot across different groups, we further design an across-group relevance constraint to balance the bias of the group priority. Experiments on three standard REC benchmarks demonstrate the effectiveness and superiority of our method.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑